Access to Education
in Five Eastern European Countries
between 1940 and 1985
Results of a Cross-National Survey"]
Harry B.G. Gauzeboom
Paul Nieuwbeerta'
This. paper presents an analysis of patterns of access to education in five Eastern European countries (Bulgaria, Czech Republic, Hungary, Russia, Slovakia) between 1940 and 1985. As in other industrial countries, average educational attainment (measured by duration, i.e. years in school) has increased greatly since 1940. But the rise was most rapid expansion until 1970. In all five countries, women increased their educational attainment more than men, By 1975, the difference between men and women in regard to years of schooling had virtually disappeared in all countries studied Parents' educational attainment (measured as average years ofschooling of father and mother) proves to be a main determinant of their children's attainment.
However, the effect ofparents' education decreased by about halfform 1940 to 1985. ellltura/resources (measured by parents' cultural behavior while the respondent was growing up) tlIrns out to be a strongpredictor of educational attainment, and equally so Jar men and women. It accountsJor about one-third ojthe educational reproduction effect. However, the effect aJcultural 'background declined considerably under communism. Finally, parents' political resources (measured as parents membership oj the Communist Party) shows to have a weak but consistent effect on educational attainment in the five countries. Although this effect was relatively strong in the early communist period, it was negligible Jor the younger cohorts.
1. INTRODUCTION
\
.
The demise of state socialism in Eastern Europe in 1989 marks the\end of what can be regarded as one of the largest-scale destratification experiments irt the history of the human race. In this experiment, one of the explicit aims of the former communist regimes was to reverse the majorfonns of social stratification and forcefully imPt?se aq equalized distribution
,
of societal rewards. To this end, the communist regimes abolished private ownership of means of production and prevented the accumulation of material and financial capital in private hands. The government instituted negative discrimination of traditionally privileged social groups (bourgeois), positive discrimination of traditionally underprivileged social groups (in particular manual workers), and strict control of migration, both within and between countries. In addition to directly influencing the distribution of scarce goods in society, marxist policyw makers • unlike marxist sociologists (Wright 1985) - have taken into account that the allocation of scarce goods does not stop at the distribution of goods within a single generation.
Italso involves the transmission of goods between consecutive generations, Marxist policyw
makers have shown an astute awareness of the fact that the family is the basic unit of social stratification. Accordingly, they knew that the transmission ofsocietal(dis)advantages between generations, e.g. from parents to their children, plays a major role in shaping the distribution of goods and opportunities. For that reason, socialist policywmakers placed the task of influencing patterns of social mobility and social reproduction high on their agenda. Specifically, influencing the intergenerational transfer of resources within families was a main target of their destratification policies. The abolition of large-scale private ownership and the
right to inherit several forms of private goods are telling examples. '
The efforts to reduce social reproduction ,went beyond policies o'n the transfer of private property and other material resources. Policies were also introduced to influence educational
selection processes.Itwas well understood that in modern societies, education serves as an
important means of transferring social inequality between generations. State socialist countries therefore actively increased the educational opportunities for children of people in lower social positions and decreased these opportunities for children of people in high positions. At some point; all communist countries took measures to enhance the educational opportunities of the' offspring of manual and farm workers and to actively discriminate against the educational advancement of others, in particular the children of (formerly) bourgeois classes. The policies adopted went far beyond the usual practices of stimulation and "head start'" programs commOn in Western societies since the 1960s. In particular, active individual discrimination was used
to select students for admission to secondary and tertiary education (or keep them out). It
would seem that these policies have to a large extent shaped the educational destinies of the
populations of Eastern· Europe, \
The first aim of this study is to assess the level of intergenerational educational reproduction in five Eastern European countries (Bulgaria, Czech Republic, Hungary, Russia and Slovakia) during the communist era. We therefore examine the extent to which the educational attainment of children is related to their parents' education. A second aim of this paper is to assess the extent to which cultural characteristics and the political status of parents mediate this relation, giving attention to 'the developments over time in the effects of these
characteristics. This paper improves on earlier workbyinvestigating data from a largewscale
survey containing information on relevant characteristics of respondents and their parents in un internationally comparable format.
2. THEORETICAL BACKGROUND AND EARLIER RESEARCH
Itis generally assumed that a person's social background has a substantial influence on his
or her educational career. In all industrialized countries, children from the lower social classes are less successful in their educational career than children from the higher classes (Shavit and Blossfeld, 1993). From their start, Marxist regimes have tried to create a more egualitarian society. Being aware of the processes of educational reproduction, they have sought to minimize the strength of the relation between education and social background. To do this, several measures have been applied in all communist countries. Especially in the 19505, strict quotas were imposed. For instance; placement in schools and universities was reserved for children of farmers and manual workers. In fact, coming from a bourgeoisbackground could
be a reason not to be admitted to higher education. Itcould therefore be expected that such
rigorous measures would have a strong impact on the distribution of educational opportunities. In the same vein, a trend towards a more open society could be foreseen. Furthermore, the chances for children from diverse social backgrounds to obtain an educational degree may be expected to become very similar. We might even expect that the chances for children from the lower social classes to get a (high) educational degree would become better than for children from the higher classes.
However, previous analyses hardly suggest that the destratification experiment has accomplished its goals in all respects, and certainly not with respect to educational opportunity (Peschar 1990, 1993). Findings on important dimensions of stratification) such as consumption patterns and occupational mobility, hardly suggest destratification across the board, as was observed some time ago (Lane 1971, 1982j Connor 1979). Inequality in access to (higher) education continues to resemble the structure of educational opportunity in other industrial societies: the social status of parents determines. the educational chances of their offspring quite strongly. There seems to have been not much change in this respect, even taken over an extended period of several decades (Simkus and Andorka 1982; Robert 1991a; Heyos and Bialeski 1990).
Several explanations have been suggested for the apparent failure of egalitarian socialist ideology or its reversal into a contradiction of its basic tenets.
One explanation of why equality policies in conununist countries did not fully live up to expectations might lie in the obstacles governments face when trying to change societies. It is probably an illusion to believe that the life chances of children from various background can easily be regulated by centrally governed policies. Even in situations calling for severe measures" peqple will always try to find a way to provide their children with a good education. And people in higher positions have more resources and probably more ambition to do so. Therefore, people in higher positions are assumed to be more successful in using or curcumventing official regulations in order to improve their own life chances and those of their children.
;.
.
\Another explanation is that the policy measures on education taken by communist regimes were not always as rigorous as suggested. According to published sources (e.g. Simkus and Andorka (1982) on Hungary; Gerber and Hout (1994) on Russia) and several spokesmen, the educational policies were upheld particularly in the 1950s, the period of orthodox
communism, but became much weaker in later decades. In Hungary, forex~mple,dedication
to these policy aims waned again from the early 1960s on, with the "backdoor" liberalization of the Hungarian economy. In Russia, such policies have never been explicit, not even at their peak in the Khrushcev era around 1960 (Gerber and Hout 1994). However, in other countries, especially Czecho·Slovakia, they were upheld much longer and lasted well into the 19805.
In the remainder of this paper, we assess the trends in intergenerational educational reproduction in Eastern Europe during the communist era. On those grounds, we can determine the relative importance of political and cultural resources for educational attainment.
In order to understand the mechanisms of educational reproduction, we have analyzed data from large-scale surveys that have been held in five Eastern European countries. Our research design uses cohorts as baseline units of historical comparison.
The data analyzed here consist of rather large-scale samples(N=about 5,000 in each country) of the general population of Bulgaria, the Czech Republic, Hungary, Russia, and Poland in 1993. The data were collected as part of the project "Social Stratification in Eastern Europe". This ongoing project is directed by Ivan Szelenyi and Donald 1. Treiman (Treiman and Some recent research has explicitly tested the hypothesis about the relatively large and growing part played by cultural resources in the stratification process under state socialist conditions; Several authors have confirmed the conclusion, reached in earlier research (see above), that socialist societies remain highly stratified in cultural respects.. However, the hypothesis that this applies to a larger extent than before has not found uneqUivocal support. For instance; in research on educational attainment in Hungary, Peschar (1990) and Ganzeboom, De Graaf and Robert (1990) find stronger effects in older cohorts than in younger ones; However, reviewing more recent Hungarian data, Robert (1991b) finds increasing effects of cultural background on educational achievement.
\
3. DATA AND RESEARCH DESIGNIfthis is indeed true, then a surprising parallel arises with the analysis made by the French sociologist Bourdieu (Bourdieu and Passeron, 1994 [1970]). He analyzed cultural strategies of social reproduction in bureaucratized; market-regulated societies. Not on the banks of the River Seine but on the banks of the Duna, Vltava, and Wisla would "cultural capital" find its best yield. According to Bourdieu's cultural reproduction theory, two basic resources determine one's position in the stratification process. On the one hand there is "economic capital", which is more than (a) monetary resources (income, wealth). It also presumes (b) economic and commercial skills, (c) a general orientation towards and taste for conspicuously displayed wealth, and (d) a corresponding social network. Bourdieu contrasts this with "cultural capital," which consists not only of (a) formal credentials and qualifications (education) and (b) knowledge and expertise, but also of (c) a general orientation towards and taste for intellectual complexity, and (d) a corresponding social network. Whereas classical stratification theories assume that material and cultural resources go together, cultural reproduction theory points to their differential consequences and their different value in varying contexts. Cultural resources are more difficult to appropriate than material resources, since they can only be generated in a long process of socialization, in the parental family and in the educational system. As a consequence, cultural inequalities are intrinsically more enduring and resistant to outside intervention than economic inequalities (Kelley and Klein 1981). Although the perverse effects of some of these exogenous processes -- such as educational expansion, the greater increased influence of parents in schools, as well as the increased credentialism in bureaucratized societies -- may have caused the importance of material resources to decline, cultural inequalities are on the rise.
Djilas (1957) still identified the members of the new class with a proletarian vanguard that entrenched its position inside the state bureaucracy; Two decaded later, Gouldner (1979) and Konrad and Szelenyi (1979) argued that the leading groups in state socialist societies are not proletarian at all, but more of an intellectual breed, (However, see Konrad and Szelenyi (1991) for critical remarks.) The latter authors have-come to characterize communist rule as the "fmal victory of intellectuals over workers (and capitalists)." The core role of education in distribution and reproduction processes of whatever kind in state socialist societies suggests they may be right. According to this argument, Plato's admonition that "the [ideal] society ... can never grow into reality' ...-till philosophers become kings in'this world, or till those we now call kings and rulers really, and truly become philosophers, and political power and philosophy come into the same hands ..." (Plato 1987 [ca. 400 B.C.]: p. 263) has become more topical than Popper (1963}would have ever imagined in his criticism of totalitarianism. Findings about the social composition of communist party members (Szelenyi 1987) show that higher educated are strongly overrepresented in the political elite.
Th~'.
idea'~ha~ ~ultural differe~ces hav~ re~la~~i:I.
theea~lier
forms of'social inequality has existed for a long time in Eastern European stratification research. Early evidence in favor of this hypothesis was provided by 'Wesolowski and; Slomczynski (1968) for Poland and Machonin (1969, 1970) for Czecho-Slovakia. These authors argued that cultural differences under socialism are particularly large and constitute the pivotal element of distribution in their societies.' This,~hypothesis has been advanced repeatedly by later Eastern European stratification researchers (e.g. Kolosi and Wnuk-Lipinski 1983; Robert 1984, 1989). Now, one may argue that this vivid interest in cultural inequality was simply a substitute for research on political and economic inequality, which was obstructed by prevailing poli.ti~al circumstances. Yet it may be true that under circumstances that prevent people fromacqulflng material resources, individuals will tend to concentrate their socially distinctive and socially reproductive behavior on the accumulation of cultural resources. Those who can successfully accomplish this will be the best off in the distribution of status and privilege.These theories of stratification processes found their best-known exponent in Djilas (1957). His "New Class'! thesis postulates that under socialism; political resources have replaced other determinants of social stratification. The ruling political class constitutes a status group with its own cultural devices to establish social closure in much the same way as envisaged by Weber (1972 [1921]) in his notion of social stratification. Accoring to Weber, the political elite ensures its own reproduction by monopolizing access to positions of power; privilege, and status. This is accomplished by direct discrimination ofothers and by exercising control over indirect channels of social mobility. If this were the case in Eastern Europe, one would expect membership of the political elite to be of major importance in gaining access to (higher) ,education. Indeed, we know that education and diplomas comprised the major selection mechanism in the stratification process under socialism.
A third ,reason why the discriminatory policies hardly had any noticeable effect in the communist countries might be that the measures had ;unintended effects. They may have created new forms of social inequality or inadvertently enlarged the differences between social groups, In this light, various "New Class" theories have been proposed by'several students of state socialist societies. These theories postulate that the traditional opposing classes (owners vs. non-owners) have been replaced by new contrasts in socialist societies.
Szelenyi 1993a; 1993b),Itis conducted in collaboration with researchers from the countries surveyed, the United States, and the Netherlands. The surveys are funded by the (American) National Science Foundation and the Netherlands Organization for Scientific Research NWO (Treiman and Szelenyi 1993a), as well as by contributions from the countries under study. The survey instrument goes beyond the usual stratification and mobility survey (e.g. Peschar 1993) in several ways. First, explicit measures of material, cultural, and political resources are used. Furthermore, the data have been collected after the transition in 1989; that is, the surveys have been held under political' circumstances that are more liberal than before. Therefore, the data can be assumed to provide a more accurate picture of peoples' educational careers and their social background. In addition, the specific contribution of the present analysis is that, unlike earlier efforts, we that we have large-scale and highly comparable data at our disposal and thereby get a general and comparative perspective on educational opportunities. Moreover, while there is much high-quality data on educational opportunities in Hungary and, to a somewhatlesser extent in Czecho-Slovakia, the datae from Bulgaria and Russia leave much to be desired. The analysis presented below should therefore be seen as a first attempt to draw these countries into the pool of countries available of the comparative analysis of social stratification and social mobility.
The surveys of these countries differ somewhat in terms of the age restrictions placed on the selected respondents,' To make the data as comparable as possible across the countries, the respondents we selected in all countries had to be between 21 and 69 years of age at the time they were interviewed. Thus, the youngest and eldest respondents were excluded. Since the data were collected in 1993, these age restrictions provide us with information onthe process of educational attainment for the several cohorts. These range from those who attended school around 1940 (our oldest respondents) to those who attended school around 1985 (our youngest respondents).
To examine trends in intergenerationaLeducational reproduction in the five countries under investigation, we have divided the data into five-year wide cohorts. These cohorts constitute the baseline units of our historical comparison. Their midpoint values correspond to the year when the members of the cohort were on average 14 years of age (That is, we coded cohort
as: cohort=year of birth
+
14. We chose this value because this is the age at which peoplemake crucial decisions on' educational careers; Moreover, this value gives the best approximation of the historical contexts (such as war and revolution) that are associated with the educational careers.Itis important to realize that by necessity, the analysis of educational attainment using cohort data refers to events that happened a long time ago for most of the respondents. The lIaveragell respondent made hislher major decision around 1960, more than 30 years before the date of survey; the oldest respondents made these decisions around 1940
and the youngest around 1985. Consequentially, the so·called "replicated cohort designll
applied in'this paper offers a unique opportunity to examine long-term trends in educational reproduction.
4. DEVELOPMENTS IN EDUCATIONAL ATTAINMENT AND PARENTS' POLITICAL AND CULTURAL RESOURCES
This paper considers the process of educational attainment over the past decades. In particular, 202
itfocuses on how parents' education, cultural reSources, and political resources affect their
children's level of education. The aim is to shed light on the developments in these effects over time. However, before we start narrowing our focus to these issues; we take another look at the big picture. We first review some developments in the level of educational attainment in the five Eastern European countries. We then trace some trends in parents' political and cultural resources.
Educational attainment is analyzed for three persons: respondent, father, and mother. As the
comparative measure of educational attainment, we use the number of years it took to achieve a given educational level. For the respondents, this was constructed from a complete roster of educational attainment. For the parents, the highest level attained was recoded into the approximate number of years of education, using expert judgment and conversions suggested by the results for the respondents.
The data were sampled in the 21-69 age range; We use the full age range for our analysis of educational attainment. In studies of educational attainment elsewhere, it is common to increase the lower age boundary. This usage is based on the argument that people in their early twenties are likely to get additional schooling, which endangers the validity of cohort comparisons for historical analysis. We prefer to remediate this problem in the design by adding (an arbitrary) two years to the score of those who are still in school. We doubt that this choice makes much difference to the final results as the large majority of that population
never gets any education in their twenties. Even among the group of21~24years old, only
a slight majority is receiving education.Itis important to look at the educational attainment
of the youngest in the datafile, as they are our only window on events since 1989. Of course, this would only apply to the small proportion of the cohort that entered tertiary education in the period of transition.
Figure 1 shows the educational distribution of men and women in the five countries in the period 1940-1985. In the beginning of that period, there were significant differences between countries in regard to the average number of years it took for people to finish their education. The mean level of education was highest (about 10 years) in the Czech Republic - especially
for men.Itwas the lowest in Bulgaria (about 6.5 years). The data also show some similarities
to other industrial countries. For instance, around 1940, the average educational attainment of women was significantly lower than that of men. Furthermore, the data show that the average educational attainment has expanded. greatly since 1940 in the socialist countries, as elsewher. But unlike other ,industrial countries, most of the rapid expansion took place in the
period 1940-1970 and was followedbya clear slowdown in the 1970s and 1980,.By1975,
Bulgarians, Hungarians, and Slovak meri and women stayed in school almost twice as long as in 1940. Educational expansion proceeded at a much slower pace for,Slovak men and for Czech men and women, all of whom started at a considerably highef. level in 1940. In particular, Czech and Slovak men have made remarkably little progress since 1940. The developments in educational attainment differ between men and women in all countries. Women have increased their educational attainment more than men. And by 1975, none of the countries displayed much difference between men and women in terms of years of schooling.
Figure1.Educational expansion ;nfive Eastern European countries, birth cohorts in school
(age 14) in 1940-1985 .
BUL CZR HUN
'::lL= c:
4050607080 - 4050607080'k::=
4050607080Figure2. Parents' Party membership in five Eastern European countries, birth cohorts in schooi (age 14) in 1940-1985 BUL CZR HUN
O'''j
~
jr
j _
?o.06iL
.(0 50 60 70 80~
40 50 60 70 80k::::
40 50 60 70 80:lst:
40 50 60 70 80 40 50 60 70 80 '"'1 AUS SLO ,c, men o women .' RUS SLOO'''j.~
j
~
O'06~ ~
.(0 50 60 70 80 40 50 60 70 80parents (mother and father) of the respondent were members ,of the Communist Party. UnfortunatelYl there is no information in our dataset on ,when parents were members of the party, nor on theiuelative positionJn, the party hierarchy,ItiS,still reasonable to assume that the questions, pick up some of the, political resources ,at the time that the respondent was growing up and major decisions about hislher educational career were made. On average, 33 percent of the respondents in the Czech Republic, 25 percent in Russia, and between 18 and 21 percent in the "remaining ,three .countries indicated that ,one of their parents, had been a member."Of. a communist political party. ,It is, -of course, ,hard to judge to-what extent respondents gave reliable reports oftheir parents',party membership. However, the percentage of respondents.-who report being a member-of a communist political party in 1993 is in accordance with' official statistics. Therefore, we, are confident that infonuation given by respondents about their parents',party membershipis,a~soaccurate..
"
The percentage of respondents in the various cohorts who reported that their parents have been members of a communist political party is given in Figure 2.,Parents' party membership was rather rare among the older cohorts, though it increased over time. The increase might be simply a period effect: the parents of older respondents had less time to become a member ofa communist party. In the Czech Republic, the percentage (36%) of parents' who had been a member of a communist party peaked around 1965. ,In the other countries,:it increased over the whole period and reached about 20 percent at its highest point.
Itshould be kept in mind that in all five countries, the group of party members included a sizable proportion of highly educated persons. The correlation between party membership and education in Russia and Bulgaria is the strongest (well over 0.30). In the other countries, this correlation is weak but consistently significant (about 0.20). These findings emphasize the advantage of examining the effects of cultural and political resources simultaneously by means of multivariate analysis.
204
To measureparents' cultural resources; the survey included a battery of questions on the cultural practices of the parents of respondents when the respondent was "around 14 years old." The items included in this questionairre closely parallel those in instruments used to measure cultural resources in other surveys. For example, the TARKI Hungarian stratification surveys of 1982 and 1986 showed strong effects of these variables on the educational attainment of respondents (Ganzeboom, De Graaf and Robert 1990). There are eight separate questions, two of which pertain to cultural behavior outside home (museums, perfonnances), while the remaining six refer to behavior at home, mostly reading habits. Of course, one may speculate on the effects of the separate components of this set of indicators. (Does reading matter more than other fonus of cultural participation?) Notwithstanding, the present paper treats these indicators as an expression of one underlying dimension: command over "cultural resources." This stance is further justified by the high intercorrelation between the eight items and the corresponding high level of Cronbach's reliability coefficient in AppendixA.The index on cultural resources was created by averaging the indicators standardized over the total sample. In order to make effects comparable cross-nationally, the resulting index was then standardized again within countries.
,
.
The means and standard deviations of the cultural resources indexf~r all the cohorts in all five countries show only moderate historical variations, which' parallels fmdings in other countries. In case of strong historical variations, it becomes an issue whether the variable should be conceptualized as a positional good that needs to be standardized within cohorts. Given the low over-time variation, this is not of much importancehere~Nevertheless, we have still taken the trouble to do this: the index of cultural resources is standardized within countries, allowing an interpretation of cultural resources as a relative (positional) good, for which the scores need to be interpreted relative to cohort means.
5. MODELS distinguished cohorts.
Figure 3. Uncontrolled effects ofparental education on respondent's education in five Eastern European countries, 'birth cohorts in school (age 14)in 1940~1985
As stated above, the aim of this paper is twofold. Let us now dwell on the first aim, which is to assess the historical variations in patterns of educational attainment. To investigate the effect of parents' education on the educational attainment of their children and the developments in that effect, we analyze our data by applying a regression model with multiplicative interaction terms. The equation for Model I , which is used in analyzing data for men and women in the five countries separately, can be written as follows:
where the variables "Child's education" and "Parents' education" (Le. the average education of father and mother) refer to the number of years it took to complete a given educational level for children and their parents, and the variable "cohort" refers to the' five-year wide cohorts we designate as units for historical comparison. The variable "cohort," as pointed out
above, originally ranged from 1940to 1985.This range covers the period during which the
oldest and the youngest respondents in our data set were 14 years of age and important
decisions on their educational careers were made. We centered this variable around the "cohort
of 1960" by subtracting1960from its original value. Accordingly, the intercept coefficient
(P2) for the effect of parents', education pertains to that effect around 1960.
Child's education= PIi"'COHORTj+
P2"':Parents' education+PJ"'COHORTj"'Parents' educaticn (1)
Second, the variable "cohort" is included in the model as an interaction tenn with the variable "Parents' education." Thenitrepresents the developments in the effect of parents' resources on their children's educational attairnnent.Itisthese developments that are under investigation here. To examine the nature of these developments, we evaluate three nested versions of this
model. In the discrete cohort model (A),itis assumed that effect of parents' education differ
between cohorts and are allowed to do so in free variation (cohortj= cohort;). We estimate
a different effect for each of the distinguished i cohorts. Since we distinguished II five-year-wide cohorts, this model requires ten degrees of freedom (number of cohorts - I) and is our
least parsimonious model. In the linear change model (B), itis assumed that the differences
between cohorts can be modeled as a linear trend (cohortj=cohortj),which requires a single
degree of freedom. Comparison between the discrete cohort model (A) and the linear change model (B) leads to a gross test of whether there are non-linearities in the historical patterns of effects. Since this is a nine degrees of freedom test, it may still hide significant spikes in the data that might be uncovered by more parsimonious specifications of the non-linearities. We guard against these possibilities by conducting a visual inspection of the parameters of models (A) and (B). Finally, we test whether the linear changes modeled in (B) are statistically significant, or whether the even simpler model (C) of no historical changes (cohortj= 0) still fits the data.
Let us now turn to the second aim of this paper and examine the extent to which the effect of parents' education on their children's education is mediated by the effects of the cultural and political resources of the parents. Concomitantly, we examine the extent to which these later effects have changed over the last decades. The model we use to study the effects and their changes Model II is similar to Model I. But the second model also included the variables for "Parents' party membership," "Parents' cultural participation" and the interaction between these variables and tlcohort" also are included:
132"'Parents' education+PJ*COHORT1"'Parents' education+
Child's education~ ~1i"'COHORTj+
",,}£'U'.
, " '.
~,~,~CZR
. . . .""1~Hu,N',",n~RUS
.
""b"O
0
" 0 0 0"
" 0 ' " ' 0
' °
0 i°
0 00,'" " " " '0
eoo,,,~
". " ""',0
~"""
" " " ,,' 0,'" " " " "
eo0,,,,,, "
eo ":0
p,,"'Parents'partymembership
+
Ps"'COHORTj"'Parents' partymembership+
P6*Parents' cult. participation+P1"'COHORT1"'Parenls' cult. participation (2)
() men o women
The'variable licohortt is included twice in Model LFirst, this variable is included as ,the intercept' and takes into account the developments in the avetage level of educational attairnnent. 'As mentioned earlier, educational expansion for both men and women in the five countries Was more progressive in the beginning of the period under investigation and weak (or even absent) near the end ofit. Therefore, we included the variable "cohort;" in a discrete
manner~Le., specifying dummy variables for the various cohorts .. allowing for these
non-linear trends. As a result, the Pli coefficient represents the mean level of education in the 206
where the variable "Parents' party membership" is coded 1for
c~ildre~\who
indicated their parents had been a member of a communist party, zero if they did not; and the variable "Parents' cultural participation" is a score on an index that we standardized within each cohort within each country: in each cohort the mean is rescaled to zero, t,he stardard deviation to 1. The centering around the mean of the scores within a cohort avoids collinearity between the variables "cultural participation" and '''cohortu• In addition, by standardizing the index of
to the effects of parents' party membership and cultural participation in the period around
1960, and the trend parameters(PsandP7)refer to absolute changes in these effects compared
to these effects in that period,
Table 1. pelected Coefficients from Regression Models for (Trends in) Effects of Social Background in Educational Attainment
Figure4. Controlled effects of(1) parents~ education (controlled for effictsof parents'
,I:', political and cultural resources) and(2)parental political resources (controlledfor
effects ofparents' education and cultural resources) on respondent's education in five Eastern European countries, birth cohorts in school (age 14) in 1940-1985
(I) political,resour~es
BUt CZR HUN FlUS StQ
""b'
"'""~"",""~",,,~ ,.,,~
, " . . OOe...~ o · 0 000' a . 0 0
.1,C8 ".tC8 a , "\,C8 0 0 000
'.',~8
a o . .1.t8 0 .co 50 60 10 80 ~o 50 60 10 eo co 50 60 10 80 ~o 50 60 10 eo ~o 50 60 10 eo
A. Educational Reproduction ca. 1960
BUL
CZR
HUN
RUS
SLO
men .400 .489 .537 .275 .560
women .467 .595 .483 .326 .600
B. Trends in Educational Reproduction (change/5 years)
men -.018 -.009n8 -.033 .005n, -.046
women -.018 -.029 -.009n, .002n8 -.047
All effects are measured in years of schooling per iJnit. Units are: Parents' Education: mean years of education father+mother. Trends: change per 5 years. ns: coefficient less than twice its standard error.
6. RESULTS
Trends in educational reproduction
BUl. " , 1 •
, .••
~.
,', ' ' ' , ' ' ' ' " ' , ,....~CZ~,'''i
" : ' , ....~HU~
' . , ' '.'" RU' 'c. '. ,',.... seo ,',.0 , " . :
,"I"·"·'
0 'b?'", b"
. 0 j., 0 " ' i 0 ' .. : 0 ,
, 00 0 .'; , ' 0 , " 0 00 0
O.19l 0.19~.0.19~ . 0194 0 0019l 0/
~50Wro~ ~50~10~ ~50W70~' ~50~70~" ~50W70~
We first address a' descriptive'. issue:' Has socialism·1changed ,the, pattern of' educational
reproduction at all? Then we discuss whether the observed patterns of change in access to (higher) education are linear or non-linear. To examine which pattern ofchange describes our data best, we employ Model I and compare three versions of it. One of these represents non-linear changes (using discrete cohorts) in the effects of parents' education on their children's
edu~ation(Model A). Another one assumes linear changes over cohorts (Model B). The third.
verSIOn assumes no ,change (Model C). The results of comparing. these models are rather
consi~te~t:
I?
al~ten cases ~.five countries, taking men and womert separately - an adequatedescnptton IS given by a Imearsummary of the developments in the effects of parents'
education on the level of education attained by their children. The relevant coefficients of
Model B for the five countries investigated are presented in Tablel,1lI ,
Let us start by examining the magnitude of the effects of parents' education on that of their children. Recall that we centered the variable "cohort" around the cohort" 196011
•Accordingly,
the effects of parents' education in the cohort "1960" are represented by the estimated
coefficients(P2) presented in Part A of Table 1. Both the dependent and independent variables
(respondent's education and the mean of father's and mother's education) in Model I are operationalized .in the same way in the five countries, and these show similar standard
deviations. Therefore! the estimated effectparameters can be compared among the countries.(ll
The coefficients vary between 0.50 and 0.60 in the Czech Republic, Hungary, and Slovakia. Internationally compared, these values are rather high (see Ganzeboom and Treiman, 1993). The values in Bulgaria (0.40 for men and 0.47 for women) are somewhat lower; but only those for Russia (0.28 for men and 0.33 for women) are somewhat low from an international perspective. On average, these results square with ·conclusions drawn in earlier studies. Namely, they suggest that communist regimes have not succeeded in truncating the intergenerational transmission of status through education. In this respect, the low values of the coefficients for the, effects of parents' education on their children's education in Russia are somewhat surprising. However, since we have no other dataset for Russia at our disposal, we can not draw definite conclusions on the (low) level of educational reproduction in Russia over the past decadesYI
The next question to be addressed is: How much have the effects ofp~rents'education on the
educational level of their children changed during the period r'940-1~85.Part B of Table 1
shows the relevant (linear) trend coefficients; which reveal that the effects of parental education have decreased in four countries for both men and 1¥0men. Again, Russia fonus an exception: the trend coefficient is virtually zero. This value indicates that hardly any long-term change in the level of educational reproduction has occurred and that Jhis effect remained at the low level discussed above during the whole period from 1940 until 1985. In two other cases (that of Czech men and Hungarian women), the trend coefficients also are statistically insignificant. However, in these cases, the trendparameters have a negative value. Not only the sign but also the value df the estimated trendcoefficients can be interpreted in a
<>
men o women (2) cultural resourcesTable 2. Selected Coefficients from Regression Models for (Trends in) Effects of Social Background in Educational Attainment
11.
Trends in Educational Reproduction (change/5 years) after controlling Parents' Culture andParty Membership
A"
Educational Reproduction ca. 1960 after controlling Parents' Culture and Party MemberShipstraightforward way, These represent a decline in the effects of parents' education per decade compared to the effects on members of the middle cohort, as presented in Part A of Table 1. Among other things, this implies that for Hungarian men and for Slovakian men and women in the last cohort (who entered the educational system around 1985), parental- influence on
their educational level was aboutone~thirdof the influence observed in the middle cohort
(who entered around 1960) and was almost as low as the Russian level. Considering the low intercept coefficient in Russia, this low Russian level is almost reached in Bulgaria even though Bulgaria has a relatively moderate negative trend coefficient. Concluding, our analysis reveals that on average the advantages ofmgher levels of education have gradually diminished over the communist period. By and large, the effects of parents' education On their children's education declined by about half over the period 1940-1985; In view of the results of earlier studies on this topic, this is an important new finding.
men women .282 .333 .299 .390 .305 .267 .156 .204 .343 .356
Effects
0/
political and cultural resourcesFinally,we ask:: To what extent were cultural resources (as measured by, parents' cultural practices when the respondent'was growing up) and political resources (as measured by parents having been a member of the socialist or communist party) important.toeducational careers under communism? The underlying issue here is the debate on intellectuals as the "New (leading) Class" under socialism (Konrad and Szelenyi 1979). To answer this question we analyze our data employing Model II , Besides parents" education; this model includes their political and cultural resources as explanatory variables.
- I'! ',."_ ': 'j ')!,~,i;l ' : ", "
Also for this model,' we' compared three versions. For each of the effect~, we assumed
respectively (A) non-linear changes across cohorts, (B) linear changes, or (C) no changes. The choice between the different versions of the models is not unequivocal for the various cases. In five'ofthe ten cases (five countries, men and women taken separately), the non-linear trend model 'describes our data best;' in the other five, the linear trend model is most suitable. However; the amount of variance explained by both models is very similar, and the more parsimonious linear version yields an adequate fit in all cases, 'Therefore, Table 2 presents the
central coefficients of the linear version of Model II. I
Parts A,C" and E' of this table 'show the (intercept)coefficients (Pl' 04' and P6)' representing the effects of parents' education;' parents"political resources, and their cultural resources,
respectively;' around 1960. Parts B,' D,' andP.show the (trend) coefficients(P3' P5; and P7)
representing the (absolute) changes per five years in the mentioned effects. Both the intercept and the trendcoefficients can be compared across countries and between men and women. The (intercept) coefficient for the effect of party membership represents the difference in the number of years of education attained between children of party members and children of parents who had not belonged toa communist party. The (intercept) coefficient for cultural resources represents the difference in years of education between children whose parents differ one standard deviation on the Z-score of the variable cultural' participation. Since we also centered the variablelIcohort" in ModelU around the'1960-cohort, the (intercept) coefficients for the effects of parents' education, cultural resourcesJ:and political resources refer to the'
effects of these factors in that period. .
210
men -.004n' .007n' -.013n' .017n, -.006ns
women -.009n' -.007n' -.OOln' .012n, -.017n'
C. Effect of Parent,' Party Membership ca. 1960
men .779 .396 .305n, .713 .374
womeri .612 .197 .062ns .707 .474
D. Trends in Effect of Parent's Party Membership (change/5 year)
men -.156 -.033n' -.I03n' -.122 -.012n'
women -.026n' -.05In' .012n' -.Olln' -.137
E. Effect of Parents' Cultural Resources ca. 1960
men ,932 .722 1.227 .596 .846
women .945 .777 1.195 .685 .933
E
Trends in Effect of Parents' Cultural Resources (change/5 years)men -.136 -.056 -.246 -.081 -.141
women -.106n' -.058 -.080 -,117 -.077
All effects are measured in years of schooling per unit. Units are: Parents' Education: mean years of
education father+mother, Parents' PartyMembership~(0) never (I) even member communist party
(father or mother). Parents's Cultural Resources: standardized index of 8 items (mean=O, stdev=l). Trends: change per 5 years. ns: coefficient less than twice its standard tITor.
-\
Aninteresting outcome of our analysis is the extent to which the effect of parents' education
on their children's educational attainment is interpreted by the added variables, namely
parents' party membership and ,cultural participation. The co~trolled,effect of parents'
mediated yiaparty membership and cultural participation. Cultural partidpationhere plays the
majorr~le, while' the effect of parents' party membership is modest.
The estlmatedcoefficients fof the effects of parents' party membership (Part C in Table 2) are all positJve' but differ substantially in size. The effect of parents' party membership, for example, is statistically insignificant in Hungary, whereas in Bulgaria and Russia these effects
are considerable (about 0.7 years). Itis probably no coincidence that these latter countries
show the weakest effects of parents' education on their children's education. However, there is evidence that the effect was relatively strong in the early communist period but had
declined toinsignificance,for·th~YOWlger cohorts.. ;, \. '
The estimated effects of parents' cultural participation ont~eirchildren's educl;ltion (Part E
in Tabk I) are not :directly comparable to the effects of the other variables. Nevertheless, since the standardized variable' has a reach of four times its standard deviation, we can conclude that this factor has a much stronger impact than parents' party membership. Without exception, ,all pertinent coefficients for parents' cultural, participation are, positive, and statistically significant. Moreover, they hardly differ between men and women. The coefficients are highest in Hungary, the country where Bourdieu'scultural reproductiontheory has been corroborated repeatedly. The smallest effect of cultural participation is. found in Russia, but even there the effect is substantial. This leads us to conclude that in all five countries, the cultural resources of parents are highly beneficial to their children's educational attairunent. This conclusion corresponds to the hypothesis that socialist intellectuals form a "New Class". Those belonging to this group are able to use the established socialist distribution of power to their own advantage.
However, the (interaction) coefficients that pertain to the trends in the effects of parents' . cultural resources on their children's education (Part F in Table 1) reveal that the effects of cultural participation have declined substantially over the past decades. Thus, if the intellectual elite took advantage of the socialist circumstances, they profited more from the orthodox than the liberal brand of socialism, The pertinent trendcoefficients have high, statistically
significant, negative values and thus imply a substantial decline in~e effects ofpar~nts'
education in most countries. For example, the effect of parents' educatIOn on the educational attainment of Hungarian men was large in the older cohorts (around 1940) but is almost absent in the most recent cohorts (around 1985): This result replicates earlier findings for the Hungarian case (Ganzeboom et,at. .1990). And this outcome implies that the decline in the effects of cultural participation, together with the decline)n the effects;.of parents' party membership, are largely responsible for the decline in the uncontrolled effects. of parents' education on that oftheir children'in the communist countries, as observed earlier.
-,"
Before making some concluding remarks, let us summarize the results of our analyses for the
various countries gr~phically.In Figure 3, ,we pre,sent tref?ds in the uncontrolled effects of
parents'
edu~at~on,
o,n that ofJheirchil~ren:
In Figure 4, we present trends in,the;~~ntrolled
effects ~fp,arents' education, aS,well as the (controlle,d) effects·~f p~rents pohtIcal and
cultural'resou~cesi\"In
these tigures, We plotted both the 'estimated values obtained from a discrete (non::-linear) 'trend ve'rsion of the models,and values from a lineartr~ndversion. Thisgives animpr~ssfqnto.w~atextent and ,in which cohorts the linearandnon-linearversion~led
to.different results,' In general, however, differences. between the outcomes of the two verSIons appear not torbe linked to certain periods. Thus, a'linear approach to the data provides an 212
adequate (and parsimonious) description of the development.
In three countries, however, the effect of parents' party membership is exceptionally high in the begiIU1ing of the period under investigation. This is curious, since the children of these cohorts were already in school when the communist regimes came to power. Thus, a relatively low effect of parents' party membership is to be expected. A possible explanation of this inconsistent finding might be that children of revolutionary commWlists benefited from their parents eliteposition later on in their (educational) career. Nevertheless, before advancing this. interpretation, we should stresS that for the oldest cohorts, the effects of party membership presented here are based upon analysis of a limited number of respondents.
7. CONCLUSIONS AND DISCUSSION
The results of the analyses discussed above lead us to three main conclusions.
Our first conclusion is that in aU five Eastern European countries under investigation, the parents' education has a substantial effect on their children's education. Although no explicit comparisons with other countries have been made, we found that the strength of the effects of parents' education in four out of the five countries is even stronger than usually found in Western countries with a market economy. Only in Russia did we find a level of intergenerational educational reproduction that is rather weak, from an international perspective.
Our second main conclUSion is that parents' party membership (to a small extent) and their cultural participation (to a considerable extent) influence the level of education their children attained. These factors explain on average about one-third to a half of the effect of parents' education on educational attainment. These findings thus support the "New Class" theories, which assume that intellectuals tend to benefit most under a state socialist regime.
The slow but consistent decline, in the detenninants of educational attainment in all countries over the last decades fonns a third important conclusion of this study. In four countries, the parents' social background lost its impact on their children's educational career; while in Russia this impact was already limited at the begirming of the period under investigation. The gradual expansion towards opeIU1ess occurred with respect both to the effects of parents' education and the effects of parents' political and cultural resources. Concluding, there seems to be an unmistakable trend toward a situation where ascriptive characteristics have only a
limited impact on educational attainment, , .',
.
.
.
.,
educational expansion and its consequences for educational reproduction remains a topic for further research.
Another
t~~ic f~r
further research'.is therema~kably
strong decline in the effects of parents'cultural resources in the five Eastern European countries. In The Netherlands, for example, recent analyses showed hardly any decline in cultural, and educational reproduction, over the past decades (Van Eijk and De Graaf 1994). Differences between Western market economies and state socialist countries might be due to the relatively strong educational expansion in Eastern Europe. However, for a better understanding of the trends in educational success and educational reproduction, both types of countries have to be compare;d explicitly. ,
NOTES
I.Earlier versions ofthis paper were presented at the: Workshop on Social Stratification in Eastern Europe after 1989, December 11-13 1994, Budapest; Meetings of the ISA Research Committee on Social Stratification at the ISA World Congress, Bielefeld, July 18-23 1994; Bi-annual Meetings of the Netherlands Sociological Association, Amsterdam, April 1994; and the Meeting orthe ISA Research Committee on Social Stratification, Canberra, January
1994. '
2. An appendix ,containing all, results ofthe analyses carried out can be obtained upon reques,t from the authors; 3. The standard deviations of the variables concerned, however, are not entirely identical; that Is, they range from 0.51 in Bulgaria to 0.32 in Russia.
4. The data set collected by Gerber and Haul (1994) also contains information on educational reproduction in the former Soyiet Union, but onlyItregional part of. that country is covered. Analysis of their data using the same models resu1t~ In ~omewhat higher rep'roduction coeffic!ents than ,Our analyses (personal communication Ted Gerber). , '
REFERENCES
Bourdieu" P., and 'J,C, Passeron" (1994[1970]); Reproduction in Education. Society and
, Culture~London: Sage. ' . . ,,
COlmor; W.O. (1979),Soc,ialism, ' Politics and Inequality. Hierarchy and Change in Eastern
'Eu'.'ope,'and the USSR. New York: C?lumbia UniversityPress;
Djilas,M,(1983[1957]), The New Class, An Analysis oJthe Communisl SySlem, San Diego:
Harcourt~ ' , ' , , . . " ,
Eijk,
K.
van,'
andP.
de Graaf(1994),"Uit
een nest?De
invloed vanprimaire socialisatie oponderwijskansen."Mens en Maatschappij (69-3). I·
Ganzeboom, H". P" de Graafand P. Robert (1990), Reproduction theory on socialist ground.
The intergeu:rational transmissi?n of status attainment in Hungary~Research in Social'
Stra(ificCJ,tion'ahd Mobility 9,,~pp; 79~104.,' rt, .', ", "
Ganzeboom, H,B.O:,;'and D,,1. Treiman(1~93),"Preliminary tesults on educational expansion' and educational ,opportunityin:'comparative' perspective','. In: H.A. Becker and P.L.],
'~7rrn;~~?~'(~~s),,!~oP~~~!7,ot~~~er~tio~s. Am~terdam:'~hesis'
Publisher,s.214
Gerber, T., and M. Hout (1994), "Educational stratification in Russia during the Soviet period." Paper presented at the World Congress of Sociology, Bielefeld, July 1994.
Gouldner, A.W. (1979),The Future a/Intellectuals and the Rise o/the New Class. London:
Macmillan.
Graaf, P. de, andH.Ganzeboom (1993), "Family background and educational attainment in
the Netherlands of 1891-1960 .birth cohorts", In: Y, Shavit and H,P, Blossfeld (eds.),
Persistent Inequality. Changing Educational Attainment in Thirteen Countries, pp.75~100. Boulder: Westview Press.
Heyns, B., and I. Bialeski (1990), "Educational inequalities in postwar Poland." Paper
presented at the ISA World Conference, Madrid.
Kelley,l, and H. Klein (1981), Revolution and the Rebirth of Inequality. A Theory Applied
to the National Revolution in Bolivia. Berkeley: University of California Press.
Kolosi, T., and E. Wnuk-Lipinski (eds.) (1983), Equality and Inequality under Socialism.
Poland and Hungary Compared. Beverly i-liIls: Sage.
Konrad, G., andI.Szelenyi (1979),The Intellectuals on the Road to Class Power. New York:
Harcourt.
Konrad, G., andI.Szelenyi (1991), Intellectuals and Domination in Post-Communist Societies.
In: P, Bourdieu and J.S, Coleman (eds,), pp. 337-361, Social Theory Jor a Changing
Society. Boulder: Westview Press.
Lane, D. (1971), The End of Inequality? Harmondsworth: Penguin.
Lane, D. (1982), The End of Social Inequality? Class, Status and Power under State
Socialism. London: Allen and Unwin.
Machonin, P. (1970), Social stratification in contemporary Czechoslovakia.American Journal
of Sociology (75),pp, 725-741.
Machonin, P. (ed.) (1969), Czechoslovak Society. Sociological Analysis of Social
Stratification~[English summary.] Bratislava: Epocha Publishing House.
Peschar,lL.(ed.) (1990),Social Reproduction in Eastern and Western Europe. Comparative
Analyses of Czechoslovakia. Hungary, the Netherlands and Poland. Nijmegen: Institute
for Applied Sociology.
Peschar, J.L. (1993), "Change and continuity in Central Europe: education and social
reproduction in comparative perspective. In: Proceedings Workshop Transformation
Processes in Eastern Europe, pp. 129-145. The Hague: NWO/SSCW/SRO.
Plato (1987[ca, 400 RC.]), The Republic [Translated by Desmond Lee, Revised Second
Edition]. London: Penguin.
Popper,K.R. (1963), The Open Society and its Enemies. London: Routledge.
Robert, P. (1984), "A multidimensional approach to social mobility". In: T. Kolosi andR.
Andorka (eds.),Stratification and Inequality, pp. 223-244. Budapest: Institute for Social
Sciences.
Robert, P. (1989), "Mobilitaetsprozesse in Ungarn unter besondere',~eruecksichtigungdef
Kulturspaere". In: Kristizna M:inicke-Gyongyosy and RalfRytl1?wskNeds),Lebensstile uns
Kulturmuster in sozialistischen Gesellschaften, pp. 142·159. Berlin.
Robert,It(1991a), "Educational transitions in Hungary from the post-war period to the end
of the 1980sOl
, European Sociological Review (7-3), pp. 213.-236.\
Robert, P. (1991 b), "The role of cultural and material resources in status attairunent process:
the Hungarian case." In:A. Toth andL.Gabor (eds.),Beyond the Great Transformation,
Shavit,
Y."and
H.P. Blossfeld (eds.) (1993), Persistent Inequality; Changing EducationalAttainment in Thirteen Countries, Boulder: Westview Press.
Simkus,A., and R" Andorka (1982), IIInequalities iIi educational attairunent in Hungary
1923-1973." American Sociological Review(47-6), pp. 740.751.
Szelenyi, S. (1987), Social Inequality and Party Membership. American Sociological Review
(52), pp. 559-573.
Treiman,DJ.,andI.Szelenyi (1993a), "Social stratification in Eastern Europe after 1989." In: Proceedings Workshop Transformation Processes in Eastern Europe,pp,I63~178.The
Hague:NWO/SSCW/SRO.
Treiman, DJ., and L Szelenyi (1993b); Social Stratification in Eastern Europe after 1989.
General Population Survey,_ [machine readable datafile.] Los Angeles: University of
California-Los Angeles.
Weber, M, (1972[1921]), Wirtschqf/ und GeseUschaft.TObingen: Mohr
Wesolowski, W., andK. Slomczynski (1968), "Social stratification in Polish cities". In: 1.
Jackson (ed.)j Social Stratification, pp. 175~211.Cambridge (UK):Cambridge ,University
Press.
Wright, E.O. (1985),Classes.London: Verso.
AppendixA: Measures of Parental Cultural Consumption in Respondent's Childhood
range means
BUL CZR HUN RUS SLO
